Student Research Work

The list below gives an overview of only some of the currently open theses. Feel free to visit us at the Robotics Research Lab to get more information.

Currently Open Theses

Title Type Description Supervisor
Changing Of Digging Method To Solve Complex Task Bachelor/Master In our further projects in the topic of excavation with heavy machines we have already implemented a solution to dig a trench automatically. Our project machine can excavate a trench with a solution which is not so precise but it can remove a big amount of mass in every trenching cycle. For some tasks it is necessary to change the approach to get for example a more precise trench. This can be reached for example by using a strategy that drags the soil away and give the trench the expected shape. Your task will be to define such further strategies for digging with different properties. Additionally you should define a mechanism to switch between the different solutions within our already defined architecture. An example could be to change from mass excavation to shaping, which can be implemented as a dragging movement.  
Find Next Trenching Start To Create A Construction Pit Bachelor/Master In our projects with the excavator we want to build a construction pit automatically. At the current state we are able to dig a trench in the range of a fixed placed excavator. To build up a whole construction pit it is necessary to dig multiple of such local trench stripes. For that we need a strategy to find at every time the next pose to place the excavator in an optimal way that it can dig the next strip. All of this stripes together should build the expected shape of the desired construction pit.  
Move An Excavator To The Next Trenching Position Bachelor In our projects with the excavator we want to build a construction pit automatically. At the current state we are able to dig a trench in the range of a fixed placed excavator. To build up a whole construction pit it is necessary to move the the machine from one excavation position to the next one. For that it is necessary to execute a driving manoeuvre which relocate the excavator to a new position. This is necessary because at a local position it only has a restricted working range.  
Intelligent Property Representation in Rough Environments Bachelor/Master An autonomous vehicle has to consider a lot of information sources, as sensor outputs, to fulfill its navigation task. This is especially true for unstructured off-road environments where conditions change permanently. Perception systems provide data about the current state of the vehicle as tire slippage, or inertial measurements. Also external data is used, which can be measured by distance sensors or cameras. There are sophisticated algorithms available which compute image segmentations, ground properties, classify obstacles and many more. A framework should be developed based on the behavior-based ib2c architecture. It should combine and evaluate different strings of information into a common representation. Based on this, the robot control should perform smart navigation decisions in unstructured environments. The approach should be tested and the system's behavior evaluated. Patrick Wolf (born Vatter)
Landmark Detection for Topological Map Construction Bachelor/Master Topological maps can be used for high level path planning through complex environments. Nodes in those maps represent landmarks which can be detected and re-recognized easily. Nodes can by also used in hybrid maps where grid maps provide details about the local area which corresponds to a respective node. Other applications are SLAM corrections that can make use of the landmark position. Your task is the selection of a suitable landmark representation and the implementation of such a detector. Based on the derived landmarks a topological map should be created online. The approach should be tested in a simulated and real-world environment. Patrick Wolf (born Vatter)
Virtual Fences for Autonomous Off-Road Vehicles Bachelor/Master Autonomous mobile robots are a recent and highly discussed topic. One of our research interests is the development of autonomous system solutions for commercial vehicles. Such vehicles often operate on construction sites or in search and rescue scenarios. An essential task is the navigation trough rough&nbsp; environments.<br />Hereby, existing mapping information is considered and taken into account for path planning. Nonetheless, a robot is often not allowed to enter specific areas in the environment to prevent endangerment workers or goods.<br />The task of this thesis is the development of a virtual fence for mobile robots. The system should use aerial maps where the working area can be defined. The robots navigation should detect blocked or forbidden&nbsp; pathways, adapt the global trajectory and prevent the robot from entering the restricted areas.<br /><br /> Patrick Wolf (born Vatter)

Recent student works one can find here.


Templates for student theses can be found here.

Finished work

The thesises, which were finished at the RRLAB can be found at their own subpages.